2 days ago
AI music with Holly Herndon, Mat Dryhurst, and Jesse Walden
Last month, in April 2023, a song that used the AI-generated voices of Drake and The Weeknd went wildly viral across social media before being taken down from streaming services for breaching copyright. Shortly after, Grimes took a more permissive approach, launching a platform to help fans create derivative music using her AI voice model, called Elf.Tech.
In the midst of these developments in AI music, Holly Herndon and Mat Dryhurst stand out as veterans who have have been on the forefront of the intersection of music, community, technology, and AI for many years. In 2021, they released a DAO-governed AI voice twin of Holly’s, called Holly+, and their more recent project, Spawning, offers tools for creators to manage their AI identities. In this conversation, we’re also joined by Jesse Walden, who originally got his start in crypto through music, by being a music manager before co-founding Mediachain, a startup which sought to attribute every piece of media on the internet using blockchains. We weave through the economics of AI music and derivative creation, the impact AI will have on the distribution of creator success, how hyper-personalization maps to listener behavior, industry regulation, impacts for record labels, and more.
Jesse and Li are cofounders and General Partners at Variant, an early-stage web3 venture firm. Learn more at https://variant.fund/
(00:00) Episode preview
(02:27) Spawning a baby and a startup
(04:07) Holly+ & AI-driven content creation
(07:02) Community participation & creation using Holly+
(11:16) Grimes & the economics of derivative works
(17:32) Crypto in the evolution of AI music
(21:03) Jesse Walden’s journey from music to crypto
(25:41) How AI impacts the power law of music success
(36:53) Music consumption modes: passive vs. active
(39:47) Hyper-personalization of music
(47:45) Updates on Spawning: AI tools for artists
(50:19) AI, data, and creator consent
(53:14) Regulation & adaptation for record labels
(58:59) Worldcoin, voice models, and digital identity
(01:03:31) Prediction for the future of music: what will stay the same?